Overclock.net › Forums › Graphics Cards › AMD/ATI › AMD MANTLE Discussion Thread
New Posts  All Forums:Forum Nav:

AMD MANTLE Discussion Thread - Page 193

post #1921 of 2769
Quote:
Originally Posted by BradleyW View Post

Quote:
Originally Posted by rickcooperjr View Post

lol just because monitor is 144hz doesnt meen you have to run 144hz run 60hz refresh rate and downscale the resolution to 1080p 1920x1080 on the game side and it is good to go and will act no different than a regular 1080p monitor. Why add more work than needed to the situation.

Actually, a 60Hz screen at 60Hz will act different from a 144Hz screen at 60Hz. The result is, 144Hz @ 60Hz is less smooth and has increased input lag + increased motion blur. It's disappointing but there we are.


He was going for recorded results via software not physical feel AKA graphs and such so the overall tests would be the same minus his end of it AKA his personal visual conclusion isnt recordable so that result doesn't matter only the software recorded results mattered in this case. I am just pointing out what he wanted and was implying the testing should be resolved with physical stats via software recorded results not his side of it. He specified physical feel meant nothing and only cared about the software recorded results to look at frame latencies and such cant do that with human eye without recorded results which is what is in question here.
Edited by rickcooperjr - 3/6/14 at 8:02am
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
post #1922 of 2769
Quote:
Originally Posted by rickcooperjr View Post

lol just because monitor is 144hz doesnt meen you have to run 144hz run 60hz refresh rate and downscale the resolution to 1080p 1920x1080 on the game side and it is good to go and will act no different than a regular 1080p monitor. Why add more work than needed to the situation.

I am not going to play BF4 at 60 hz, like a peasant!

post #1923 of 2769
Quote:
Originally Posted by the9quad View Post

Quote:
Originally Posted by rickcooperjr View Post




you do realize can run your 1440p monitor and just run it at 1080p at 60hz for testing purposes no need to completely change the monitor and remember have to fully remove other cards otherwise Mantle gets confused so just disabling them via device manager or CCC is a no go.

lol yes rick i know that. but it's right here on my desk and it does 144hz, this one doesnt at 1080p. Thanks though seriously, you never know when I will do something bone headed.


I also understand you wish to put that screen to use at its full potential and proper settings but for testing results sometimes you have to vary your settings to be more inline with other testing platforms to test variances. I want to say the changing monitors alot can be bad and increases the chance of damaging like say ports on graphics card or monitor or increases chances of dropping or scratching monitor so by just mimicking the test specs via settings will bring you to exact conclusion as changing the monitor itself minus your end of it graphics wise lol as the term goes think smarter not harder also remember this patience is a virtue but when doing testing the amount of time used testing various settings and such can be alot of time so the areas you can cut that wont add extra time and yet will get same results so if can avoid altering hardware ( increases chances of damage and also adds alot of time and often hassle ) and instead alter settings if possible do this unless no other option.


By the way I like your term of refusing to play BF4 like a peasant lol but facts are not many use 144hz screens or 1440p dominant testing range will be 60hz 1080p so trying to compare your results to others is like comparing a ferrari to a 73 F250 high boy truck ferrari is faster and costs alot more but F250 highboy can crush the ferrari then drag it across the muddy field and even possibly bury its carcass lol again variance all depends on usage / settings so very possible your tests are not relevant to 90% of players of BF4 or users of Mantle because so few use 144hz screens or 1440p. To sum it up basically best option is to mimmick general population of users hardware ( AKA monitor res and refresh rate ) to get a good idea of general performance of Mantle not throw something at it that is so top tier that it is irrelevant to most of users then complain about the discrepencies in the results via the mainstream users.


I want to point out myself I am not a mainstream user I run 3x 32inch 2560x1600 or 1600p 60hz refresh rate screens and often do 7680x1600 eyefinity so my results are not mainstream nor relevant to general population or users of Mantle.
Edited by rickcooperjr - 3/6/14 at 8:38am
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
post #1924 of 2769
Anyway, re-installed 14.2's over the 14.2's, much better but as you can see still a hitch at the end. would like to see some more people post their results with longer benches, as you can see I don't get the hitch until after 25,000 frames.:


Edited by the9quad - 3/6/14 at 8:19am
post #1925 of 2769
Is that with single 290x or CF? Ill try and have DX vs Mantle charts for my setup tonight. But for sure Mantle outperforms DX11 at 1440p High (no AA) for me.
Edited by jerrolds - 3/6/14 at 8:37am
post #1926 of 2769
3 cards in crossfire.

another one, notice it doesn't happen until after 40,000 frames here, but once it starts it happens more often.



So I am still not convinced this just doesn't happen to everyone. As you can see all the other people have not posted runs this long. The only one I saw so far was < 20,000 frames, and as you can see mine are butter smooth until after 20 or 40 thousand frames also.
Edited by the9quad - 3/6/14 at 8:41am
post #1927 of 2769
Quote:
Originally Posted by the9quad View Post

3 CARDS IN CROSSFIRE. If you becnh long enough I'm pretty sure your graphs will start showing hitching though.

another one, notice it doesn't happen until after 40,000 frames here:


That is interesting if you can run GPU z or somethign to monitor the Vram usage also look for paging file issues bump paging file on hardrive to around 12gb or more I noticed anything under 8gb of paging file even with 16gb 4x 4gb 1866mhz ram BF4 would act weird without enough paging file size I am running 2x 290X's so this might be relevant to you also to maybe resolve some of the issue.
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
post #1928 of 2769
Quote:
Originally Posted by rickcooperjr View Post

That is interesting if you can run GPU z or somethign to monitor the Vram usage also look for paging file issues bump paging file on hardrive to around 12gb or more I noticed anything under 8gb of paging file even with 16gb 4x 4gb 1866mhz ram BF4 would act weird without enough paging file size I am running 2x 290X's so this might be relevant to you also to maybe resolve some of the issue.

I dont think page file will help, as in when I use directx 11 as the renderer this doesn't happen. So it's definitely something with Mantle and how it is doing something ( i am thinking some kind of vram/memory leak something or other). Which means you can probably run for a really long time at 1080p before it happens or a really short time at 1600p or higher.
Edited by the9quad - 3/6/14 at 8:54am
post #1929 of 2769
Quote:
Originally Posted by the9quad View Post

Quote:
Originally Posted by rickcooperjr View Post

That is interesting if you can run GPU z or somethign to monitor the Vram usage also look for paging file issues bump paging file on hardrive to around 12gb or more I noticed anything under 8gb of paging file even with 16gb 4x 4gb 1866mhz ram BF4 would act weird without enough paging file size I am running 2x 290X's so this might be relevant to you also to maybe resolve some of the issue.

I dont think page file will help, as in direct 11 this doesnt happen.

seriously try it set it to 12gb and see what happens at anything under 8gb on Mantle i get all kinds of weird stuff going on I bumped it to 8 is good with a few issues bumped to 12gb and all is good 90% of time I was exactly like you didn't understand how paging file could affect it but it does Mantle is new and different so all previous understanding goes to the wind so try it and retest it remember after upping paging file size to 12gb need to restart computer for it to take affect.

I want to point out I had so much issues running at 1600p till I upped paging file size to 12GB then 90% of my Mantle issues left for me running 2x 290X's in crossfire on BF4 I dont understand why it fixxed main issues or how it did but it did so PLZ give it a try is 1minute or so to change paging file size to 12+gb ( general rule of thumb is 2-3 times amount of ram you have ) and a system reboot then retest your results not back breaking or insanely complicated to do.
Edited by rickcooperjr - 3/6/14 at 9:05am
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
Intel I7 5820k 4.6ghz core 4.3ghz CPU cache 1.345v asus sabertooth X99 sold R9 290x's am currently running a R9 280A X... 16gb of ddr4 corsair vengeance LPX 3200mhz 1t 1... 
Hard DriveOptical DriveCoolingOS
M.2 Toshiba OCZ RD400 512gb ASUS DVD rw  liquid cooled cpu and gpu's windows 10 pro 64bit 
MonitorKeyboardPowerCase
3 x 32lcd/led 2560x1600 wolfking timberwolf EVGA supernova 1300w G2  Thermaltake core X9 
MouseMouse PadOther
steel series wow cataclysm 14button programable  wolfking sniper Intel PRO PT dual port PCIe server NIC 
  hide details  
Reply
post #1930 of 2769
Quote:
Originally Posted by rickcooperjr View Post

seriously try it set it to 12gb and see what happens at anything under 8gb on Mantle i get all kinds of weird stuff going on I bumped it to 8 is good with a few issues bumped to 12gb and all is good 90% of time I was exactly like you didn't understand how paging file could affect it but it does Mantle is new and different so all previous understanding goes to the wind so try it and retest it remember after upping paging file size to 12gb need to restart computer for it to take affect.

I want to point out I had so much issues running at 1600p till I upped paging file size to 12GB then 90% of my Mantle issues left for me running 2x 290X's in crossfire on BF4 I dont understand why it fixxed main issues or how it did but it did so PLZ give it a try is 1minute or so to change paging file size to 12+gb ( general rule of thumb is 2-3 times amount of ram you have ) and a system reboot then retest your results not back breaking or insanely complicated to do.

I was at 32gb, since i have 32 gigs of ram and 32 gigs seemed large already lol. Think about it the more memory you have the less likely you are to page data out, so the more ram you have the less valid that 2 to 3 times your memory suggestion is.


You recommend I lower it to 12GB?
Edited by the9quad - 3/6/14 at 9:09am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › AMD MANTLE Discussion Thread